US7756034B2

US7756034B2 - System and method for handling network overload

Info

Publication number: US7756034B2
Application number: US11/290,711
Authority: US
Inventors: Steven R. Donovan; Ben A. Campbell; Adam B. Roach; Robert J. Sparks; Ajay P. Deo
Original assignee: Cisco Technology Inc
Current assignee: Cisco Technology Inc
Priority date: 2005-11-29
Filing date: 2005-11-29
Publication date: 2010-07-13
Also published as: US20070121515A1

Abstract

A system and method for handling network overload includes receiving one or more requests, wherein an originating client originates the one or more requests. It is determined if a network element handling the one or more requests is overloaded. If the network element is overloaded, a pushback message associated with a specific request of the one or more requests is generated. The pushback message is sent to the originating client.

Description

CROSS-REFERENCE TO RELATED APPLICATION

This application is related to U.S. application Ser. No. 11/290,997, filed on Nov. 29, 2005 by Steven R. Donovan, et al. and entitled “System and Method for Handling Network Overload.”

TECHNICAL FIELD OF THE INVENTION

This invention relates generally to the field of communications and more specifically to a system and method for handling network overload.

BACKGROUND

Networks facilitate communication between clients using network elements, which may be arranged in clusters. Conventional methods for handling overload in a network element include ignoring the possibility of overload or discarding requests from a client when the network element becomes full.

Another conventional method includes balancing the traffic load between network elements in the cluster, which causes the elements in the cluster to approach overload at approximately the same time. When the element reaches overload, conventional methods include returning a pushback response to be handled by the previous element. The pushback message results in the previous element in the network not sending requests to the overloaded element for a period of time. During this time, traffic diverts to other elements in the cluster. Because the other elements in the cluster continue handling their same traffic load, the additional traffic from the overloaded element causes other elements in the cluster to go into an overloaded state also. Consequently, a portion of the network goes into an overloaded state, which prevents the flow of traffic between clients.

SUMMARY OF THE DISCLOSURE

From the foregoing, it may be appreciated by those skilled in the art that a need has arisen for an improved system and method for handling network overload. In accordance with the present invention, disadvantages and problems associated with conventional systems and methods to handle network overload may be reduced or eliminated.

According to one embodiment of the present invention, a system and method for handling network overload includes receiving one or more requests, wherein an originating client originates the one or more requests. It is determined if a network element is overloaded. If the network element handling the one more requests is overloaded, a pushback message associated with a specific request of the one or more requests is generated. The pushback message is sent to the originating client.

Certain embodiments of the invention may provide one or more technical advantages. A technical advantage of one embodiment includes using an end-to-end message to transmit a pushback message from the overloaded network element to the client. By communicating directly to the client instead of using intervening network elements to process the pushback message, the portion of the distributed network with the overloaded network element may continue to facilitate communication. Another technical advantage of another embodiment includes providing request gapping or retry-after information in the pushback message. Therefore, the remainder of the distributed network may continue to operate without experiencing a shutdown. Yet another technical advantage of another embodiment includes allowing the client to choose an action in response to the pushback message. For example, the client may choose to send the request to another network element located in a different cluster. The different cluster may handle a different level of traffic and allow the request from the client to pass through without encountering an overloaded network element.

Certain embodiments of the invention may include none, some, or all of the above technical advantages. One or more other technical advantages may be readily apparent to one skilled in the art from the figures, descriptions, and claims included herein.

A non-transitory computer readable medium including logic for handling network overload, the logic operable to: receive one or more requests, wherein an originating client originates the one or more requests; determine if a network element handling the one or more requests is overloaded; generate a pushback message associated with a specific request of the one or more requests if the network element is overloaded; send the pushback message to the originating client.

BRIEF DESCRIPTION OF THE DRAWINGS

For a more complete understanding of the present invention and its features and advantages, reference is now made to the following description, taken in conjunction with the accompanying drawings, wherein like reference numerals represent like parts, in which:

FIG. 1 illustrates a network that includes network elements to route requests made by clients;

FIG. 2 illustrates an overloaded network element sending a pushback message toward the client;

FIG. 3 is a flowchart illustrating a method for sending the pushback message toward the client;

FIG. 4 illustrates an overloaded network element sending a pushback message with request gapping information;

FIG. 5 is a flowchart illustrating a method for sending the pushback message with request gapping information;

FIG. 6 is a flowchart illustrating a method for receiving the pushback message at the previous hop with request gapping information.

DETAILED DESCRIPTION OF THE DRAWINGS

FIG. 1 illustrates a network 10 that includes network elements to route requests on behalf of clients 100. Network 10 includes network elements, such as

servers

106 and 108, that facilitate communication between clients 100 by routing the signaling or control requests of clients 100.

Clients

100 exchange audio, voice, data, video, or other information in network 10. To control the exchange of the information, clients 100 send signaling or any suitable control request within network 10 to establish communication with other clients 100. The requests may include any suitable information, such as a message, and the request may be in any suitable communication protocol. Clients 100 may be any combination of hardware and/or software that provide communication services to a user. Client 100 may include analog, digital, or Internet Protocol (IP) telephones, a cellular phone, a wireless fidelity phone, a personal computer with a telephony application, a personal digital assistant, or any other suitable communication device. Although network 10 illustrates a particular number of clients 100, any number and arrangement of clients 100 and network elements is contemplated.

Network 10 allows clients 100 to communicate with each other. Network 10 may facilitate communication using messages in any suitable protocol, such as Session Initiation Protocol (SIP) or Signaling System Number 7 (SS7) protocol. In an embodiment, network 10 is a distributed network that distributes client responsibility among different network elements, such as an Internet Multimedia Subsystem (IMS), a SS7 network, or any suitable distributed network. Network 10 may include any combination of servers and any other hardware and/or software.

Network elements are included in a Point of Presence (PoP) 102. PoP 102 contains a variety of network elements and may interact with other PoPs 102. Network 10 may include any suitable number of PoPs 102. Each client 100 may be assigned to a particular PoP 102. For example, client 100 a may use PoP 102 a and PoP 102 b to communicate with client 100 b, while client 100 c uses PoP 102 c and PoP 102 d to communicate with client 100 d.

In the illustrated embodiment, each PoP 102 includes separate clusters 104 of

servers

106 and 108. Clusters 104 include multiple network elements, and may separate the network elements by type.

Servers

106 and 108 may be any suitable server type, such as proxy servers, back-to-back user agent servers, or any suitable combination of servers. For example, cluster 104 a includes any suitable number of servers 106, which are proxy servers, while cluster 104 b includes any suitable number of servers 108, which are back-to-back user agent servers.

Servers

106 participate in routing requests from an originating client 100 to a terminating client 100. In an embodiment, servers 106 are proxy servers. Each server 106 may represent a logical entity that transmits requests, messages, or other communication. Server 106 may handle any suitable number of requests at any suitable rate. Within network 10, servers 106 are the entry and exit points for requests of client 100. In an embodiment, servers 106 are responsible for load balancing requests among servers 108 to provide uniform distribution of traffic within PoP 102. Server 106 may support any suitable communication protocol, and may implement any suitable feature related to request routing, such as authentication, authorization, and compression. For example, if network 10 uses SIP, server 106 may be an edge proxy, whereas if network 10 is an IMS network, server 106 may be a Proxy CSCF (P_CSCF).

Servers

108 participate in routing any suitable type of request between clients 100. In an embodiment, servers 108 are back-to-back user agents. In yet another embodiment, servers 108 are proxy servers. In addition to routing requests, servers 108 coordinate the traffic for clients 100 served by PoP 102. In an embodiment, servers 108 load balance the requests among servers 108 in another PoP 102 or among servers 106 in PoP 102. Servers 108 may include any suitable combination or arrangement of logic operating to support routing requests between clients 100. Server 108 may support any suitable communication protocol. For example, if network 10 uses IMS, server 108 is a serving CSCF (S_CSCF).

In operation, an originating client 100 a sends a request to server 106 in PoP 102 a assigned to client 100 a. Server 106 a routes the request to server 108 a in the same PoP 102 a. Server 108 a routes the request to server 108 b in PoP 102 b associated with a terminating client 100 b. Server 108 b routes the request to server 106 b, which routes the request to the terminating client 100 b. Because the request is routed by each network element, a hop in network 10, from originating client 100 a to terminating client 100 b, the request is routed hop-by-hop through network 10.

Each request from each originating client 100 a goes through network 10 hop-by-hop to reach terminating client 100 b. Servers 108 may become overloaded while handling the received requests. For example, server 108 becomes overloaded if it receives more requests than it can handle, if the requests come faster than expected, if the traffic is congested, or if any other suitable occurrence happens. When server 108 reaches an overloaded state, the portion of network 10 that includes server 108 may continue to operate if server 108 uses one of the following mechanisms to handle the overload: pushing back to the client or pushing back hop-by-hop using request gapping.

Modifications, additions, or omissions may be made to network 10 without departing from the scope of the invention. For example, PoP 102 may include any suitable number of network elements and clusters. As another example, network 10 may facilitate communication between clients 100 using any suitable type of network element. Additionally, operations in network 10 may be performed using any suitable logic.

FIG. 2 illustrates an overloaded network element sending a pushback message 206 toward client 100. Client 100 a sends request 200 to server 106 a in its associated PoP 102 a. Server 106 a

routes request

200 to server 108 a at message 202. Server 108 a

routes request

200 to server 108 b at message 204. Server 108 b is in PoP 102 b that is associated with terminating client 100 b. In the illustrated embodiment, server 108 b has reached an overloaded state. Accordingly, server 108 b generates a pushback message 206 to inform client 100 a of its overloaded state. Server 108 b sends pushback message 206 toward client 100 a. Server 108 b may send pushback message 206 directly to client 100 a or route pushback message 206 to client 100 a through intervening

servers

106 and 108, as shown in the illustrated embodiment. If pushback message 206 is routed through intervening

servers

106 and 108, the intervening

servers

106 and 108 do not process or handle pushback message 206.

Pushback message

206 is a response from server 108 to request 200 sent from client 100 a. Pushback message 206 may be generated by any suitable server, such as a SIP server, an IMS server, or a SS7-enabled server. Pushback message 206 may include any suitable information regarding the overloaded state of server 108. For example, pushback message 206 may include retry-after information that provides a time interval in which client 100 may re-send the request. Client 100 may wait for the indicated time interval to expire, alert the user of the condition of server 108, or send request 200 to another server 108. In an embodiment, server 108 sends pushback message 206 with retry-after information for a single user client 100.

As another example, pushback message 206 may include request gapping information. Request gapping information provides client 100 with details about sending requests 200, such as a minimum interval between requests or a maximum frequency to send requests. For example, the request gapping information may inform client 100 to send no more than 100 messages every second, a message no faster than every 10 milliseconds, or any suitable frequency, duration, or interval. When client 100 receives pushback message 206, client 100 may throttle the sending of requests 200 according to the request gapping parameters received in pushback message 206. In an embodiment, server 108 sends pushback message 206 with request gapping information when server 108 can originate requests, such as if server 108 is a conferencing server. Request gapping information is more fully discussed with reference to FIG. 4.

FIG. 3 is a flowchart illustrating a method for sending pushback message 206 toward client 100. At step 300, server 108 receives request 200 from client 100, which has been routed through network 10 using other network elements, such as

other servers

106 and 108. At step 302, it is determined whether server 108 is in an overloaded state. If server 108 is in an overloaded state, it generates pushback message 206 at step 304. Server 108 sends pushback message 206 to originating client 100 at step 306. Client 100 may receive pushback message 206 directly from server 108 or pushback message 206 may be routed through intervening

servers

106 and 108. If pushback message 206 is routed through intervening

servers

106 and 108,

server

106 and 108 do not process or handle pushback message 206. As discussed with reference to FIG. 2, pushback message 206 may include any suitable information, such as retry-after information or request gapping information. The method returns to step 300 after server 108 sends pushback message 206 to originating client 100. The method continues until server 106 is not in an overloaded state.

If server 106 is not in an overloaded state, server 108 routes request 200 toward terminating client 100 at step 308. When routing request 200 toward terminating client 100, request 200 may proceed through other network elements before reaching terminating client 100. The method subsequently ends when request 200 is routed toward terminating client 100.

FIG. 4 illustrates an overloaded network element sending a pushback message 406 to a previous hop with request gapping information. Client 100 a sends request 400 to server 106 a in its associated PoP 102 a. Server 106 a routes the request to server 108 a at message 402. Server 108 a routes the request to server 108 b at message 404. In the illustrated embodiment, server 108 b has reached an overloaded state. Accordingly, server 108 b generates a pushback message 406 to inform client 100 a of its overloaded state. In an embodiment, pushback message 406 is a SIP message, such as a 503 Service Unavailable response.

Pushback message

406 includes request gapping information. Request gapping information provides server 108 a with details about sending request 400, such as a minimum interval between requests or a maximum frequency to send requests. The request gapping information may be included as new parameters within an existing header, such as a Retry-After header, or as a new header. The request gapping information may include any suitable parameters, such as duration, interval, frequency, or any suitable combination of parameters. Duration refers to the duration that request gapping should be applied. Interval refers to the minimum time server 108 must wait between sending requests 400. Frequency refers to the maximum requests server 108 may handle per unit of time. Frequency is an alternative parameter to the interval parameter.

Request gapping information may apply to any suitable type of request 400. For example, the request gapping information may apply to all types of incoming requests 400, or the request gapping information may apply to specific types of requests 400. In an embodiment, server 106 may pushback lower priority requests 400 with request gapping information and may allow higher priority requests 400 to flow without gapping.

When server 108 a receives pushback message 406, server 108 a may throttle the sending of requests 400 to server 108 b by exercising request gapping options. Request gapping options may include any suitable action to discontinue sending requests 400 to server 108 b. For example, a request gapping option may include routing pushback message 406 to another upstream element. In the illustrated embodiment, server 108 a

routes pushback message

406 to server 106 a in message 408. Another request gapping option may include server 108 a

queuing request

400 for later transmission consistent with the request gapping information. Yet another request gapping option includes server 108 a

routing request

400 to another downstream server, such as server 108 c.

FIG. 5 is a flowchart illustrating a method for sending pushback message 406 to a previous hop with request gapping information. At step 500, server 108 b receives request 400 from client 100, which has been routed through network 10 hop-by-hop using other network elements, such as server 106 a and server 108 a. At step 502, it is determined whether server 108 b is in an overloaded state. If server 108 b is in an overloaded state, server 108 b generates pushback message 406 at step 504. Server 108 b sends pushback message 406 with request gapping information to the previous hop at step 506, which is server 108 a in the illustrated embodiment. The method returns to step 500 after server 108 sends pushback message 406 to originating client 100. The method continues until server 108 b is not in an overloaded state. When server 108 b is not in an overloaded state, server 108 b routes request 400 toward terminating client 100 at step 508. When routing request 400 toward terminating client 100, request 400 may proceed through other network elements, such as server 106 before reaching terminating client 100. The method subsequently ends when request 400 is routed toward terminating client 100.

FIG. 6 is a flowchart illustrating a method for receiving the pushback message at the previous hop with request gapping information. At step 600, server 108 a may exercise request gapping options. As discussed with reference to FIG. 4, request gapping options may include server 108 a

routing pushback message

406 upstream through network 10 hop-by-hop, re-routing request 400 to another downstream server 108, or queuing request 400 for later transmission consistent with the request gapping information.

At step 602, it is determined whether request 400 has been re-routed to another downstream server 108. If server 108 a

re-routes request

400 at step 602, the method subsequently ends. If server 108 a does not re-route request 400 to another server at step 602, request 400 is rejected at step 604. The method subsequently ends if request 400 is rejected.

The flowcharts are only exemplary illustrations. Modifications, additions, or omissions may be made to the flowcharts. In addition, steps and messages may be performed in any suitable manner.

While this disclosure has been described in terms of certain embodiments and generally associated methods, alterations and permutations of the embodiment and methods will be apparent to those skilled in the art. Accordingly, the above description of example embodiments does not constrain this disclosure. Other changes, substitutions, and alterations are also possible without departing from the scope and spirit of this disclosure.

Claims

1. A method for handling network overload, comprising:

receiving one or more requests at a network element from one or more intervening network elements, wherein an originating client originates the one or more requests to establish communication with a terminating client;

determining, at the network element, if the network element handling the one or more requests is overloaded;

generating a pushback message, at the network element, associated with a specific request of the one or more requests if the network element is overloaded, wherein the pushback message includes information indicating a time for transmitting subsequent requests; and

sending the pushback message from the overloaded network element to the originating client for the originating client to process, wherein the network element, the originating client, the terminating client, and the one or more intervening network elements are separate devices and the one or more intervening network elements do not process the pushback message, and wherein sending the pushback message to the originating client includes sending the pushback message directly to the originating client.

2. The method of claim 1, wherein sending the pushback message to the originating client includes routing the pushback message associated with the specific request from the overloaded network element to the originating client through one or more network elements that routed the specific request to the overloaded network element.

3. The method of claim 1, wherein the information indicating a time for transmitting subsequent requests includes request gapping information, the request gapping information including a selected one of a minimum interval to receive the one or more requests or a maximum frequency to receive the one or more requests.

4. The method of claim 1, wherein the information indicating a time for transmitting subsequent requests includes retry-after information, the retry-after information including a time interval to receive a request from the originating client.

5. The method of claim 4, further comprising receiving the one or more requests a second time from the originating client in accordance with the retry-after information.

6. The method of claim 1, further comprising providing an alert to the originating client if the network element is overloaded.

7. The method of claim 1, further comprising routing the one or more received requests to a terminating client if the network element is not overloaded.

8. The method of claim 1, further comprising:

determining a priority of the one or more received requests;

generating the pushback message according to the priority of the one or more received requests.

9. The method of claim 1, wherein generating a pushback message includes generating a session initiation protocol (SIP) message.

10. Logic encoded in a non-transitory computer readable medium for handling network overload, the logic when executed operable to:

receive one or more requests at a network element from one or more intervening network elements, wherein an originating client originates the one or more requests to establish communication with a terminating client;

determine, at the network element, if the network element handling the one or more requests is overloaded;

generate a pushback message, at the network element, associated with a specific request of the one or more requests if the network element is overloaded, wherein the pushback message includes information indicating a time for transmitting subsequent requests; and

send the pushback message from the overloaded network element to the originating client for the originating client to process, wherein the network element, the originating client, the terminating client, and the one or more intervening network elements are separate devices and the one or more intervening network elements do not process the pushback message, and wherein sending the pushback message to the originating client includes sending the pushback message directly to the originating client.

11. The non-transitory computer readable medium of claim 10, wherein sending the pushback message to the originating client includes routing the pushback message associated with the specific request from the overloaded network element to the originating client through one or more network elements that routed the specific request to the overloaded network element.

12. The non-transitory computer readable medium of claim 10, wherein the information indicating a time for transmitting subsequent requests includes request gapping information, the request gapping information including a selected one of a minimum interval to receive the one or more requests or a maximum frequency to receive the one or more requests.

13. The non-transitory computer readable medium of claim 10, wherein the information indicating a time for transmitting subsequent requests includes retry-after information, the retry-after information including a time interval to receive a request from the originating client.

14. The non-transitory computer readable medium of claim 13, the logic further operable to receive the one or more requests a second time from the originating client in accordance with the retry-after information.

15. The non-transitory computer readable medium of claim 10, the logic further operable to provide an alert to the originating client if the network element is overloaded.

16. The non-transitory computer readable medium of claim 10, the logic further operable to route the one or more received requests to a terminating client if the network element is not overloaded.

17. The non-transitory computer readable medium of claim 10, the logic further operable to:

determine a priority of the one or more received requests;

generate the pushback message according to the priority of the one or more received requests.

18. The non-transitory computer readable medium of claim 10, wherein generating a pushback message includes generating a session initiation protocol (SIP) message.

19. A system for handling network overload, comprising:

one or more clients operable to send and receive one or more requests, wherein an originating client sends the one or more requests and processes a pushback message and a terminating client receives the one or more requests from the originating client;

one or more network elements operable to route the one or more requests, a network element receiving the one or more requests of the originating client from one or more intervening network elements, the network element operable to:

determine, at the network element, if the network clement handling the one or more requests is overloaded;

send the pushback message, from the overloaded network element, directly to the originating client for the originating client to process, wherein the network element, the originating client, the terminating client, and the one or more intervening network elements are separate devices and the one or more intervening network elements do not process the pushback message, and wherein sending the pushback message to the originating client includes sending the pushback message directly to the originating client.

20. The system of claim 19, wherein sending the pushback message to the originating client includes routing the pushback message associated with the specific request from the overloaded network element to the originating client through one or more network elements that routed the specific request to the overloaded network element.

21. The system of claim 19, wherein the information indicating a time for transmitting subsequent requests includes request gapping information, the request gapping information including a selected one of a minimum interval to receive the one or more requests or a maximum frequency to receive the one or more requests.

22. The system of claim 19, wherein the information indicating a time for transmitting subsequent requests includes retry-after information, the retry-after information including a time interval to receive a request from the originating client.

23. The system of claim 22, the overloaded network element operable to receive the one or more requests a second time from the originating client in accordance with the retry-after information.

24. The system of claim 19, the overloaded network element operable to provide an alert to the originating client.

25. The system of claim 19, the one or more network element operable to route the one or more requests to the terminating client if the network element is not overloaded.

26. The system of claim 19, the network element operable to:

determine a priority of the one or more received requests;

27. The system of claim 19, wherein generating a pushback message includes generating a session initiation protocol (SIP) message.

28. A system for handling network overload, comprising:

means for receiving one or more requests at a network element from one or more intervening network elements, wherein an originating client originates the one or more requests to establish communication with a terminating client;

means for determining, at the network element, if the network element handling the one or more requests is overloaded;

means for generating a pushback message, at the network element, associated with a specific request of the one or more requests if the network element is overloaded, wherein the pushback message includes information indicating a time for transmitting subsequent requests; and

means for sending the pushback message from the overloaded network element directly to the originating client for the originating client to process, wherein the network element, the originating client, the terminating client, and the one or more intervening network elements are separate devices and the one or more intervening network elements do not process the pushback message, and wherein means for sending the pushback message to the originating client includes means for sending the pushback message directly to the originating client.

29. A method for handling network overload, comprising:

receiving one or more requests, wherein an originating client originates the one or more requests to establish communication with a terminating client;

determining, at a network element, if the network element handling the one or more requests is overloaded;

determining a priority of the one or more received requests;

generating a pushback message, at the network element, associated with a specific request of the one or more requests if the network element is overloaded, wherein the pushback message includes information indicating a time for transmitting subsequent requests, and wherein the pushback message is generated according to the priority of the one or more received requests;

sending the pushback message from the overloaded network element to the originating client, wherein the network element, the originating client, the terminating client, and one or more intervening network elements are separate devices and the one or more intervening network elements do not process the pushback message, and wherein sending the pushback message to the originating client includes sending the pushback message directly to the originating client.